Performance-Based Mix Design for Porous Friction Courses (2018)

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

35 Introduction The work plan was divided into two parts. Part 1 was to evaluate mix design parameters, performance tests, and performance criteria that can be included in a performance-based speci- fication. Part 2 was to further evaluate the effect of specific mix components on the PFC mix performance parameters determined in Part 1. Six different mix designs were included in Part 1, and two of these mix designs were included in Part 2. The six mix designs were selected based on their field performance and mineralogy. Three aggregate types (granite, limestone, and traprock) were used in this study. Three poorly perform- ing mix designs were selected from Florida (limestone), South Carolina (granite), and Virginia (traprock), while three well performing mix designs were selected from Florida (limestone), Georgia (granite), and New Jersey (traprock). Based on agency comments, the well performing mixes have had services lives up to 19 years before being replaced, while the poorly performing mixes were replaced within 10 years. An exception was the Florida good performing mix only had a service life of 9 years at the time this study began, but it had 50% more traffic than the mix considered as a poor performing mix. Table 10 shows a summary of the project information. Projects for each aggregate type are located in similar environments and weather conditions. The Florida projects are even within the same county and used the same aggregate source. Similar asphalt content and binder type were used except for the Florida project, where one had ground tire rubber (GTR) modifier and the other had SBS binder modifier. The difference in binder content for the Florida projects is primarily due to the addition of rubber to the GTR (ARB-12) mix. The two granite projects are from different quarries, but used the same material specifications, design method (pie plate), same binder grade, and same construction requirements. However, this research study shows there is a definite difference in mix performance and verifies that you can have two mixes meeting the same specifications that have very different results. It emphasizes the need for performance testing to be required for PFC mixes. The two traprock mixes were designated by the agencies as 9.5 mm PFC. However, the Virginia mix had 86% passing the 9.5 mm sieve, which would technically classify it as a 12.5 mm NMAS mix. The Virginia mix had 10% RAP with PG 82-22 rubber modified binder. The stiff binder with the addition of RAP may have made the Virginia mix too stiff for the cold environment. It was the northernmost project in the state and had the most winter maintenance activity. The New Jersey 9.5 mm mix used PG 76-22 binder with rubber modifier. In order to make findings from the study applicable to states that do not use hydrated lime as an anti-strip agent in PFC mixes, the research team agreed that a liquid anti-strip agent would be used for all mix designs evaluated in this study. In addition, for PFC mix designs with hydrated C H A P T E R 3 Work Plan

36 Performance-Based Mix Design of Porous Friction Courses lime, the lime would be replaced by baghouse fines to maintain the total dust content. The liquid anti-strip agent used in all the mixes in this study is added to the binder at a dosage of 0.5% by weight of the binder and blended before mixing with aggregate. Verification of PFC Mix Designs The six mix designs were verified in the laboratory before testing was conducted. The job mix formulas (JMFs) for each of the mixtures tested can be found in Appendix B. All of the aggregate specimens were taken from the same quarry referenced in the JMFs. Gradations of the aggregate materials sampled for this study were performed, and they were slightly different from those in the JMFs. This was expected as some of these mix designs were done over 20 years ago. The variation in production or sampling may cause the difference between the historical and current gradations. To account for the difference, the aggregates were fractionated and then batched so that the aggregate gradations of the mixtures tested in this study were as close as possible to those in the respective mix designs. Specific gravity testing was conducted on all of the stockpiles as well so that the voids in mineral aggregate (VMA), voids in the coarse aggregate (VCA), and the film thickness could be calculated accurately. Part 1—Development of Performance-Based PFC Mix Design Currently, there are two national standard mix design procedures for PFC: ASTM D7064 and AASHTO PP 77. These procedures are similar to each other and include requirements for selecting materials, design gradation, optimum binder content, and performance testing for draindown, raveling, permeability, and moisture susceptibility. In Part 1, a performance-based mix design procedure was to be developed based on these procedures. Table 11 shows important parameters in four areas, including selecting materials, determin- ing the aggregate structure, determining the optimum asphalt content, and conducting perfor- mance tests, for a performance-based PFC mix design procedure. These important parameters were initially considered as a baseline for producing a well performing PFC mix design and were therefore determined for each of the six mix designs. Results of Part 1 were then analyzed to determine which parameters and associated criteria may be related to PFC mix performance in the field. The performance tests were conducted in Part 1 based on the following test procedures: • AASHTO T 305—Draindown Characteristics in Uncompacted Asphalt Mixtures, • AASHTO T 283 (modified)—Tensile Strength Ratio, State Route Traffic (AADT) Aggregate Type Binder Type Age, yrs. Performance (Failure Mode) FL I-95 240,000 Limestone ARB-12 9 Good (Cracking) FL I-75 160,000 Limestone 76-22 SBS 8 Poor (Cracking) GA I-85 200,000 Granite 76-22 SBS 19 Good (Raveling) SC I -20 30,000 Granite 76-22 SBS 7 Poor (Raveling) NJ I-195 40,000 Traprock GTR 76-22 19 Good (Raveling) VA 7/15 Bypass, Leesburg 50,000 Traprock GTR 82-22 3 Poor (Raveling) AADT = average annual daily traffic. Table 10. Summary of project information.

Work Plan 37 • AASHTO T 324—Hamburg Wheel-Track Testing, • ISSA TB 100—Wet Track Abrasion Test, • Tex-248-F—Overlay Test (OT), • AASHTO TP 108—Cantabro Loss, and • Illinois Test Procedure 405—Determining the Fracture Potential of Asphalt Mixtures Using the Illinois Flexibility Index Test (I-FIT). Part 2—Optimizing Performance-Based Mix Design Procedure for PFC The experiment in Part 2 was designed to evaluate the contribution of filler, binder modifi- cation, fiber, and thickness-to-NMAS ratio to the resistance of PFC mixtures to raveling and cracking. Part 2 included three experiments. A description of each experiment follows. Experiment 1—Effect of Added Dust The original approach was to use the good granite mix design and alter the dust content by adding 3.0% and 6.0% dust to the existing mix design to determine if durability could be improved. The stockpile percentages were altered for both of these options to keep the blend gradation as close as possible to the JMF. After some preliminary testing on design specimens, at varying ACs, it was concluded that 3.0% and 6.0% added dust decreased the air void content below acceptable limits. The 6.0% added dust produced an average permeability value of 11 meters/day and an average air void content of 11.4%. Since these results were unacceptable for a PFC mix design, the added dust content was changed to 2.0% and 4.0%. The 4.0% added dust produced • Materials Selection • Optimum Asphalt Content (AC) • LA Abrasion: ≤ 30% • FAA: ≥ 45% • F&E: ≤ 10% @ 5:1 • SE: ≥ 50 • PG: 1 to 2 grades higher • Anti-strip: liquid • Fiber: cellulose • NDES: 50 gyrations • % air void (Va): ≥15% • %Pb: min. 6% (VCAmix ≤ VCAdrc) • %VMA: minimum point on curve • Film Thickness: ≥ 24 µm • Passes all performance requirements below • Gradation Optimization • Performance Requirements • 3 trial gradations • #200: min. 2%, max. 8% • VCAmix ≤ VCAdrc • Draindown: max. 0.3% (2.36 mm wire basket) • Raveling: • Cantabro loss (max.): 20% (unconditioned) and 30% conditioned (freeze–thaw) • Wet Track Abrasion: Criteria not yet developed • Stripping (TSR): ≥0.7 ; psi: ≥50 • Hamburg Stripping Inflection Point: ≥5,000 cycles (after freeze–thaw conditioning) • Permeability: (min. 100 m/day) • Cracking: • OT: Min. 200 cycles • I-FIT: FI of 8.0 or greater Note: F&E = flat and elongated; SE = sand equivalency. Table 11. Expected critical factors for a PFC performance-based mix design.

38 Performance-Based Mix Design of Porous Friction Courses an average permeability value of 66 meters/day and an average air void content of 14.3%. While this was deemed close to acceptable based on the anticipated 15% minimum air void content, it was decided that this mix was already practically optimized and was only slightly improved when 2.0% dust was added to increase durability. For this reason the poor granite mix design was included in the testing plan. This additional testing allowed for the comparison of a good and a poor mix with added dust. Data supporting these decisions and explanations can be found in Chapter 5. The testing plan for the granite designs with 2.0% and 4.0% added dust is shown in Table 12. Performance tests, shown in Table 12, were conducted in this experiment to determine the mix resistance to raveling and cracking. Experiment 2—Evaluation of Binder Modification This experiment was designed to determine what effect binder, and its modifications, had on the performance of the mix. The test plan for this experiment is shown in Table 13. The Georgia good mix design was used for this part of the experiment and was tested without adding more dust to the mix. The PG 76-22 binder modified with SBS was tested with and without cellulose fiber (0.4%), while the other two binder grades were tested without fiber. This was done to determine if fibers were necessary with modified binders. Several performance tests were conducted in this experiment to evaluate the mix resistance to raveling and cracking. The PG 76-22 binder modified with GTR was blended in the laboratory prior to mixing the performance specimens. A PG 67-22 binder was used as the base grade, and a minus 30 mesh GTR was added at a rate of 12.0% by weight of the virgin binder. A heating mantle was used to Mix Design Added Baghouse Fines (BHF) Cellulose Fiber Binder Performance Test Georgia Granite “Good” 2% 0.4% PG 76-22 (SBS) Tensile Strength Ratio Cantabro Hamburg I-FIT Overlay Test (OT) Permeability Draindown Wet Track Abrasion 4% South Carolina Granite “Poor” 2% 0.3% 4% Table 12. Testing plan for Experiment 1. Mix Design Binder Cellulose Fiber Performance Test Georgia Granite “Good” PG 76-22 (SBS) 0.0% & 0.4% Tensile Strength Ratio Cantabro Hamburg I-FIT OT Permeability Draindown Wet Track Abrasion PG 76-22 (GTR) 0.0% PG 82-22 (Highly Modified) 0.0% Table 13. Testing plan for Experiment 2.

Work Plan 39 keep the binder at the appropriate temperature while blending the GTR. A high shear paddle mixer was used at a rate of 700 rpm when adding the GTR. The GTR was added to the virgin binder over the course of 2 minutes, and then the binder was continually blended for an addi- tional 30 minutes at 1,000 rpm. Care was taken to keep the binder from exceeding a temperature of 325°F while blending. The blended binder was divided into quart cans after blending due to concern that the GTR may settle in a larger container. After dividing the binder into quart cans, the binder was stirred with a glass rod prior to mixing every specimen to ensure it was a repre- sentative sample of the modified binder. The highly modified asphalt (HiMA) was prepared by the asphalt binder supplier by starting with a base binder that had a low temperature grade of -28°C. It was then modified with SBS polymer to achieve a final grade of PG 82-22. The PG 82-22 binder contains approximately 7.5% SBS polymer, which is roughly double the rate of polymer used in a typical PG 76-22 binder. Experiment 3—Effect of Lift Thickness-to-NMAS Ratio The purpose of this experiment was to determine what effect the ratio of lift thickness-to- NMAS had on the performance of the mix. Table 14 shows the testing plan for Experiment 3. Three mixes with different NMASs were evaluated. Two of the designs were PFC mixes, and the other was a dense-graded mix. The first PFC mix was the Georgia 12.5-mm granite mixture from Part 1, and the second PFC mix was a 9.5-mm PFC mixture tested in the 2012 NCAT Test Track experiment. The dense-graded mix was a 4.75-mm mix design from another study at NCAT. The mix designs for each of these mixes can be found in the Appendix. To determine the effect of lift thickness-to-NMAS on the performance of the mix, specimens were compacted in a gyratory compactor to their design height for the PFC mixtures and to a height of 95 mm for the dense-graded mixture. They were then cut into three different lift thicknesses and were saturated, frozen, and thawed (AASHTO T 283) prior to testing for indirect tensile strength. The 4.75-mm dense-graded mix was chosen to determine the sensitivity of the test results to change in layer thickness. If the tensile strength of the dense-graded mix is not affected by the varying specimen thicknesses, then this test plan will show that the difference in tensile strengths of the porous mixes may be due to aggregate structure and cohesion and not lift thickness. If the different lift thicknesses for the dense-graded mix show varying tensile strengths, then this strategy may not be applicable in evaluating the durability of the mix in terms of the relationship of lift thickness-to-NMAS. Mix Design NMAS Layer Thickness Performance Test Georgia Granite PFC 12.5 mm 1. 0.75 in. 2. 2.5 in. 3. 2.5 x NMAS Splitting Tensile Strength NCAT Test Track Section E9-1A PFC 9.5 mm Dense-Graded Mix – Lee Road 159 4.75 mm Table 14. Testing plan for Experiment 3.